A Theory of Multiple-Source Adaptation with Limited Target Labeled Data.
Yishay MansourMehryar MohriJae RoAnanda Theertha SureshKe WuPublished in: AISTATS (2021)
Keyphrases
- target domain
- labeled data
- domain adaptation
- multiple sources
- transfer learning
- unlabeled data
- cross domain
- semi supervised
- semi supervised learning
- training data
- labeled training data
- active learning
- text classification
- semi supervised classification
- test data
- training samples
- supervised learning
- domain specific
- labeled examples
- data points
- labeled and unlabeled data
- class labels
- learning algorithm
- co training
- database
- d objects
- text categorization
- prior knowledge
- training set
- label propagation
- training and test data
- training examples
- viewpoint
- object recognition
- decision trees
- information retrieval